The classification of the trivial source modules in blocks with cyclic defect groups (Q2041926)

In modular representation theory of finite groups, one of the high points is the theory of \(p\)-blocks with cyclic defect groups due to R. Brauer and E.C. Dade, where \(p\) is a prime. A quite a few part of representation theory can be viewed as kind of generalizations of parts of the cyclic block theory in a sense. Actually, the paper under review concerns a \(p\)-block \(B\) of a finite group \(G\) with a (non-trivial) cyclic defect group \(D\). More precisely, the authors look at the (so-called) Brauer tree \(\mathcal T:=\mathcal T(B)\) of \(B\), that is defined by the \(p\)-decomposition matrix of \(B\), thanks to the celebrated work of \textit{E.C. Dade} [Ann. Math. (2) 84, 20--48 (1966; Zbl 0163.27202)]. From the Brauer tree \(\mathcal T\), all (finitely generated) indecomposable modules are independently constructed by \textit{G. J. Janusz} [Ann. Math. (2) 89, 209--241 (1969; Zbl 0197.02302)] and \textit{H. Kupisch} [Math. Z. 108, 77--104 (1969; Zbl 0188.09002)]. On the other hand, in general it is well known that every indecomposable trivial source \(kG\)-module \(M\) uniquely lifts to a trivial source \(\mathcal OG\)-lattice \(\hat M\), that means that \(M\cong \hat M\otimes_{\mathcal O}k\) as \(kG\)-modules where \(\mathcal O\) is a complete discrete valuation ring of characteristic zero whose residue field \(\mathcal O/\operatorname{rad}\mathcal O\) is \(k\) that is of characteristic \(p\) (and we assume that \(k\) and \(\mathcal O\) are big enough to discuss representation theory for \(G\) and the defect group of the \(p\)-block of \(G\) to which \(M\) belongs is not necessarily cyclic). In fact, the first author and \textit{N. Naehrig} [Bull. Lond. Math. Soc. 44, No. 5, 974--980 (2012; Zbl 1259.20013)] look at the indecomposable liftable modules living in \(B\) in case that the defect group \(D\) of \(B\) is cyclic. Actually, the authors of the paper under review completely classify all (indecomposable) trivial source \(kG\)-modules in \(B\) where \(B\) is assumed to have a cyclic defect group \(D\) by making use of the Janusz classification of indecomposable \(kG\)-modules in \(B\). The way the authors take in order to get to the goal is a standard/natural/canonical one that goes back to Dade [loc. cit.]. That is, firstly one starts with the Brauer corresponding block \(\beta\) of \(B\) living at the level of \(C_G(D_1)\) where \(D_1\) is the unique subgroup of \(D\) of order \(p\). Surely, \(\beta\) is a nilpotent block, so that the Brauer tree \(\mathcal T(\beta)\) has only one edge. And then via the inertial group of \(\beta\) one climbs to the next place \(b_1\) that is the Brauer correspondent of \(B\) at the level of \(N_G(D_1)\). The Brauer tree \(\mathcal T(b_1)\) is still simple, namely, its shape is a star with the exceptional vertex in the center. Then one has to climb from \(b_1\) to our final goal \(B\). The last way is the most crucial and interesting part of cyclic block theory, though it is known that \(B\) and \(b_1\) are in fact (so-called) stably equivalent as \(k\)-algebras. One can use the Green correspondence from \(b_1\) to \(B\). It is impossible to discuss/describe one of the main theorems Theorem 5.3 because it has quite a big list. Another main result of the paper under review is Theorem 5.1, that describes the location of each non-projective indecomposable trivial source \(kG\)-module \(M\) belonging to \(B\) in the stable Auslander-Reiten quiver \(\Gamma_s(B)\), in terms of the distance between \(M\) and the boundary. This is also of much interest. Anyway, the paper is clearly well written.